nixos/systemd: enable systemd cgroup accounting by default

[flokli]

Motivation for this change

#29095 (comment)

fixes #29095.

Things done

• Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS)
• Built on platform(s)
  • NixOS
  • macOS
  • other Linux distributions
• Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
• Tested compilation of all pkgs that depend on this change using nix-shell -p nix-review --run "nix-review wip"
• Tested execution of all binary files (usually in ./result/bin/)
• Determined the impact on package closure size (by running nix path-info -S before and after)
• Ensured that relevant documentation is up to date
• Fits CONTRIBUTING.md.

Notify maintainers

cc @davidak @arianvp

[abbradar]

Looks good! I'll test it in the coming week and merge.

[davidak]

@garbas said in 2018:

I think having it false by default is ok, we can have a different discussion about it if we want to enable this or not and when.

Source: #36772 (comment)

So here we are.

For me, it's OK to have it just enabled when needed (e.g. for netdata). I havn't used it outside that.

But since Ubuntu, Fedora and Arch Linux have it enabled by default, we could enable it too by default for everyone. Or just desktop users.

Source: https://docs.netdata.cloud/collectors/cgroups.plugin/#monitoring-systemd-services

And since systemd now enables some options by default, we should have at least enabled them. And why not all to provide the full feature to our users.

Source: #29095 (comment)

Like said earlier, i had no performance issues, but i didn't measure the difference.

[davidak]

I tested it using the provided tests and can verify that the option and tests work!

[mmahut]

@GrahamcOfBorg test systemd

[mmahut]

@flokli looks like you have added this test section as well?

machine# sysctl: cannot stat /proc/sys/net/core/default_qdisc: No such file or directory
machine: exit status 1```

[flokli]

Hm - should be part of another PR (which is already merged). It probably sneaked in here as well - will have a look later today.

[flokli]

@mmahut I rebased on latest master. The diff in the tests should look clean now.

There's still another failure in the systemd test:

error: command `dumpe2fs /dev/vdb | grep -q "^Last mount time: *n/a"' unexpectedly succeeded at /nix/store/39fb7m572abixj1qi7ddzx5lp3k64lw2-nixos-test-driver/lib/perl5/site_perl/Mach
ine.pm line 376, <__ANONIO__> line 194.                                                                                                                                               
(12.12 seconds)  

But that's unrelated to this PR - and the cgroup accounting specific tests pass.

[mmahut]

@flokli thank you, have a conflict in the release-notes, do you mind to rebase, please?

[flokli]

@mmahut done - fixed the indentation of the previous changelog entry too.

[mmahut]

@GrahamcOfBorg test systemd

[flokli]

@mmahut the systemd test is still broken in master (as written in the comment above), but the newly added subtest does work.

If we're fine with this PR, we should merge it.

I opened #67555 to track the systemd vm test breakage on master.